Generative AI tools, which entered the software development scene in 2021 with the launch of GitHub Copilot, have significantly altered the landscape of open-source projects. According to a Linux Foundation study, these tools allow developers to cut their workload in half, enabling them to reinvest time into learning, new features, or addressing technical debt.
However, this shift has created a substantial burden for project maintainers. As developers increasingly rely on AI assistants, the volume of incoming code submissions has surged. Maintainers now face the challenge of filtering out low-quality contributions, with 25% of all new code submissions identified as AI-generated junk.
The issue is particularly acute in large-scale projects, where the proportion of low-quality AI-generated commits rises to 31%. While 59% of maintainers report a growing AI presence in their projects, this figure climbs to 69% for larger initiatives. Despite these challenges, only 35% of open-source projects have established formal policies regarding AI usage.
The study highlights a discrepancy in how AI is utilized: while 61% of developers claim to perform thorough testing and review of AI-generated code before submission, many others send it without any revision. Consequently, 21% of respondents noted that AI usage can actually increase the time required to complete complex tasks due to the need for extensive corrections.
